Changes in EEG frequency characteristics during sevoflurane general anesthesia: feature extraction by variational mode decomposition

Abstract

Abstract Purpose: Mode decomposition is a method of extracting the characteristic intrinsic mode function (IMF) from various multidimensional time-series signals. Variational mode decomposition (VMD) searches for IMFs by optimizing the bandwidth to a narrow band with the norm while preserving the online estimated central frequency. In this study, we applied VMD to the electroencephalogram (EEG) analysis during general anesthesia. Methods: Using a bispectral index monitor, EEGs were recorded for 10 adult surgical patients (mean age ± sd: 43.7 ± 18.5 years) who were anesthetized with sevoflurane. We created an application named EEG Mode Decompositor, which decomposes the recorded EEG into IMFs and displays the Hilbert spectrogram. Results: During the 30-minute recovery from general anesthesia, the mean of the bispectral index increased from 45.4 ± 6.8 to 96.7 ± 1.5, whereas the central frequencies of IMF-1, IMF-2, IMF-3, IMF-4, IMF-5, and IMF-6 increased significantly from 0.35 ± 0.21 Hz to 0.18 ± 0.11 Hz, 1.54 ± 0.73 Hz to 5.32 ± 3.94 Hz, 5.79 ± 2.31 Hz to 15.64 ± 8.02 Hz, 10.34 ± 2.75 Hz to 26.13 ± 7.89 Hz, 14.23 ± 3.63 Hz to 35.50 ± 4.78 Hz, and 13.74 ± 5.42 Hz to 43.20 ± 3.16 Hz, respectively. Conclusion: The characteristic frequency component changes in the specific IMFs during emergence from general anesthesia were visually captured. EEG analysis by VMD is useful for extracting distinct changes in an EEG during general anesthesia.